Commits

Commits on Oct 22, 2024

Add out_dtype kw argument to optimize_bsr_dense_addmm (#136626)

pearu
authored and
pytorchmergebot
committed

Commits on Sep 25, 2024

Tune bsr_dense_addmm for int8 inputs on A100 (#136088)

pearu
authored and
pytorchmergebot
committed

Commits on Sep 16, 2024

Add scaling arguments to bsr_dense_addmm (#136104)

pearu
authored and
pytorchmergebot
committed

Commits on Aug 22, 2024

Tune _int_bsr_dense_addmm for int8 inputs on A100 (#134035)

pearu
authored and
pytorchmergebot
committed

Commits on Aug 21, 2024

Add int8 support to bsr_dense_addmm and bsr_dense_mm Triton kernels (#133855)

pearu
authored and
pytorchmergebot
committed

Commits on Aug 5, 2024

Add tests to bsr_dense_addmm_meta. Tune bsr_dense_addmm kernel for ViT shapes. (#132646)

pearu
authored and
pytorchmergebot
committed

Commits on Aug 2, 2024

Add pinned memory support to sparse COO/CSR/CSC/BSR/BSC tensors (#129645)

pearu
authored and
pytorchmergebot
committed

Commits on May 9, 2024

Fix sparse fake tensors detach (#125679)

pearu
authored and
pytorchmergebot
committed

Commits on May 1, 2024

Require nnz==0 in sparse meta tensors (#125221)

pearu
authored and
pytorchmergebot
committed

Commits on Apr 30, 2024

Fix hybrid sparse COO tensor conversion to meta tensor (#125120)

pearu
authored and
pytorchmergebot
committed

Commits on Apr 24, 2024

Fix a bug in retrieving approximate bsr_dense_addmm kernel meta data (#124371)

pearu
authored and
pytorchmergebot
committed

Commits on Apr 18, 2024

Add index_reduce decomposition (#122579)

pearu
authored and
pytorchmergebot
committed

Commits on Apr 17, 2024

Fix index_reduce sampler filter when op_info.variant_test_name is specified (#123375)

pearu
authored and
pytorchmergebot
committed

Commits on Apr 6, 2024

Add sparse COO/CSR/CSC/BSR/BSC meta tensor input support to torch.sum (#121673)

pearu
authored and
pytorchmergebot
committed

Commits on Apr 3, 2024

Fix zeros_like on sparse compressed fake tensors (#123084)

pearu
authored and
pytorchmergebot
committed

Commits on Apr 2, 2024

[BE] Add torch.ops.aten._sparse_compressed_tensor_with_dims (#123083)

pearu
authored and
pytorchmergebot
committed

Commits on Mar 23, 2024

Update bsr_dense_addmm kernel parameters for sizes 3 x 2 ^ N (#122506)

pearu
authored and
pytorchmergebot
committed

Commits on Mar 4, 2024

Add sparse compressed fake tensor support (#120920)

pearu
authored and
pytorchmergebot
committed
Add a decomposition for torch.put, 2. (#120179)

pearu
authored and
pytorchmergebot
committed

Commits on Mar 1, 2024

Add sparse compressed meta tensor support (#120707)

pearu
authored and
pytorchmergebot
committed
Make SparseCsr a functionality dispatch key (#120703)

pearu
authored and
pytorchmergebot
committed

Commits on Feb 25, 2024

Add meta device support to sparse compressed tensors (#120498)

pearu
authored and
pytorchmergebot
committed

Commits on Feb 11, 2024

Add lowerings to special functions (#119187)

pearu
authored and
pytorchmergebot
committed

Commits on Feb 8, 2024

Add lowering to special.modified_bessel_i0 (#118993)

pearu
authored and
pytorchmergebot
committed

Commits on Feb 2, 2024

Add lowering to special.bessel_j1 (#118992)

pearu
authored and
pytorchmergebot
committed

Commits on Jan 30, 2024

Add lowering to special.bessel_j0 (2nd try) (#118565)

pearu
authored and
pytorchmergebot
committed

Commits on Jan 10, 2024

Add at::sparse::full_coo_indices utility function. (#116352)

pearu
authored and
pytorchmergebot
committed

Commits on Jan 7, 2024

Add batched sparse CSR/CSC/BSR/BSC to sparse COO conversion support (#116206)

pearu
authored and
pytorchmergebot
committed

Commits on Dec 22, 2023

Fix torch.detach doc-string (#115850)

pearu
authored and
pytorchmergebot
committed

Commits on Dec 20, 2023

Fix sparse compressed tensor invariants checks when nnz==0 (#115826)

pearu
authored and
pytorchmergebot
committed

Commits on Dec 17, 2023

SymInt'ify sparse_compressed_tensor (#107903)

pearu
authored and
pytorchmergebot
committed
Fix a crash in sparse compressed tensor invariants check when nnz == 0 (#115825)

pearu
authored and
pytorchmergebot
committed

Commits on Dec 14, 2023

Add values backward support for sparse CSR, CSC, BSR, and BSC tensors (#115586)

pearu
authored and
pytorchmergebot
committed

Commits on Dec 13, 2023

[BC breaking] Remove check_sparse_nnz argument of gradcheck (#115658)

pearu
authored and
pytorchmergebot
committed

Commits on Dec 12, 2023

Fix a fast mode gradcheck bug where specified eps argument is ignored when switching to slow mode (#115634)

pearu
authored and
pytorchmergebot
committed